I need some technical advice here.

In 2008 we had a situation where some people posted in doppler, while others used trackpoints.
This led to problems...

I requested tracks from various people on a semi-random basis, and without fail, every time there was a problem, it was when that person was using trackpoint data. The issue I have is that many people aren't all that experienced at figuring out which points are valid, and which are spikes. And spikes seem to happen far more to trackpoint data than doppler.

Garmins also played a large part in the problems we had with people claiming a speed that we didn't agree with.

So for 2009, I'd like to level the playing field somewhat. I don't want people to feel that they are being disadvantaged by using doppler... so I want to implement the 'claimed speed' that Manfred Fu chs and others on the GPS-SS site have been talking about.

This is quite easy to do from a technical perspective, just put in some 'fudge factors' for the various divisions, which relates to the type of GPS you're currently using. So if you use a garmin then the fudge factor for the 2 second speed might be -1.5, or if it's doppler on a GT-31 it might be -0.05 (or something, I don't know the exact numbers!).

I can implement this so that it's really easy to see, most people probably won't notice. And it should make the competition fairer, and hopefully avoid the temptation to use garmin data when NAVI data is available. I don't want anyone to feel like leaving the challenge


How does everyone feel about this? Technically, what numbers should I use? Would this system discourage people from posting?

Nebbs, do all the software programs give claim speeds?
i am using gpspro, i am not sure if it has this feature?

One thing though how can guaranty that the Software is setup the same for each person.

Perhaps we need a default rsp file for realspeed that everyone uses for this challenge im not sure if those settings are in the rsp file though.
maybe Timemachine can answer this.

ta Vando

To clarify how this would work:

When inputting data, you still type exactly the same numbers as you would have done so before, but you also select which GPS (and trackpoint/doppler) you used.

The website would then adjust your scores according to the type and accuracy of the GPS that you have selected.

Personally, I'd leave it as is.

It is to hard and complex to calculate an accurate "Fudge factor".

If you eliminate old GPS's then some people may not want to or afford to bother replacing them.

If you want have a box to tick as to what equipment you are using.

As Hardie said Trackpoints only comes out better for the small distances the larger ones give better on doppler.

I would have a clause about mixing TP and Dop for the one session (Alpha excluded) in case there are some real sad cases out there (not saying there are mind you) who run both so as pick and chose between settings.

Still yet to see the cattle stations on the prize list.

A levelish playing field is good enough if we are not to uptight about it.

If people demand more there is allways GPSSS or organised events.

This is suposedly all done for Fun.

KISS

...we need to make it an even playing field...

umm.... why? The GPSTC isn't playing for "sheep stations" as has been said so many times. By using multiple categories, point-scoring and averaging two sailors, the error correction is insigificant - even for the 2sec category. Seriously, what field are we trying to level?

GPSSS already provides users with the defacto standard ranking for individuals - if Nebs wants to enhance the GPSTC website specifically for claimed speeds, then fine.... say that, but dont claim it as being better for the team aspect.

Note that the "claimed speed" adjustment doesn't fix the problem of spikes not being detected -> they are unreleated issues, eg: if there is a problem with a garmin-track (as indicated), then subtracting 1.5kn from its value wont help.

Hardie, the science here is that *statistically* we know that a given model of GPS has a certain amount of inaccuracy for a given datapoint. Over a longer period of time, that inaccuracy varies sometimes reducing, sometimes increasing the accuracy of the speed for that division. By using many tracks it becomes possible to determine the adjustment needed for say 99.5% confidence that the sailor did *at least* the claimed speed (probably more, but we cant tell due to gps inaccuracies of each datapoint). For reference, existing ratification techniques require the sailor to go *at least* the error value faster than the previously recorded value.

What will be the fudge factor for my older GPS. It's a GT-001 Chip Log my grandfather gave to me

I'm finding it really hard to download the log-line data to my commodore 64 and keeping the sand from getting wet in the hour glass also - I suppose it could go in an aquapac.

hardie said...

..................................

Doppler gives higher speeds for NM and 1 hr, and I think distance?
.............................

I have not seen any examples of this that I can remember. I have always found ALL Trackpoint results from Navi's slightly faster than Doppler results, even in the NM and I hour. If anyone has a track with an example of this I would be keen to examine and analyse it.

Realspeed has the facility built in to produce 'claimed speeds' for the various categories. GPS-Results did have that in earlier versions but Manfred seems to have removed it from the later versions. I am sure it would be a simple matter to put it back in if there was a request for it.
I have to sheepishly admit that GPSAR has got me beaten. I just can't seem to get it to do the things I want it to do in the user interface, apart from just getting a simple result. I don't know if GPSAR Pro has the facility to display 'claimed speeds' as I haven't used it much for quite a while.

I think it may be a bit more complex than Ben anticipates to do it on the web side. The deductions for each division/ catagory is a different value. It is actually very well determined now thanks to the great statistical analysis work by Manfred. There is no need for a 'Fudge' factor at all. See here: http://www.gps-speed.com/gpsaccuracy.html It is probably much better to have the values plugged into the analysis programs and have people just post the 'claimed speeds'.

Yes, it would definitely affect the Garmin users more, especially in the 2 sec short time periods. It is the ideal technical solution though, and if the Garmin users were OK to agree to it it would be a good idea, and a great incentive to upgrade to a GT-31 sooner than you may have otherwise. If we were to do this it would definitely make sense to go over to Doppler for GT-11/31's, probably for everything including Alphas as I think the GT-31 results for Alphas with the latest firmware are looking very good. Alpha Racing is the one area where we might need a bit more statistical analysis to get a better handle on the 'correction' factors for the various types of GPS but that should not take too long.

Regarding errors in posts: This is more of an issue with Trackpoints and definitely more of an issue with Garmins because the data is not in the files for the programs to use in the automatic error (spike) filters. With Navis and use of Doppler, (or even trackpoints for that matter) errors in posts should be very rare indeed if the perameters are correctly set up. (That might be something that needs more attention.)

I will get some pointers and examples together in a post to help with analysing Alphas as soon as I have some time.

sausage said...

What will be the fudge factor for my older GPS. It's a GT-001 Chip Log my grandfather gave to me

I'm finding it really hard to download the log-line data to my commodore 64 and keeping the sand from getting wet in the hour glass also - I suppose it could go in an aquapac.

Hi Sausage. Love that quaint technology!
The first deduction is to take out any Knots in the line. They will definitely give a false reading.. Also, it must be a bit of a pain winding it all up at the end of each run!

Nebs,my vote is to keep it the same with no adjustment. As we take an average of two sailors to get a speed any difference is haved.(Does that make scence) Lets keep posting in track points with all its problems,then were comparing apples with apples. If people want the absolute speeds then post to GPSSS as well,but people shouldn,t compare the 2 sites as they are different totaly comps. As to spikes in tracks ,a thread or article written to help all people through the various programs on how to get the right info out of your gps would be great as not every has a Daffy or Matt to help. I like that anyone can post with any GPS, we need more people sailing and having fun.

I don't care who posts with what gps. I'm racing myself.

Hey I'm glad this thread is up because i was going to ask a very related question.
Being new to GTC i am using Garmin 201 and i have noticed that my 2sec peaks in realspeed are ALWAYS lower than my peak speed readout on the Garmin. For example today i peaked at 30.7 on the garmin but only 29.04 on my 2 sec peak. that just doesn't seem right.

so do i assume correctly that there appears to be a 'lowside' error with garmin when posting numbers to realspeed?. if so, i would like to see some simple form of adjustment or i'll have to buy a navi, because it is frustrating me big time when i go and sail what i think is a fast session next to others to then see much lower peaks compared to others posts (that's not to say that i should be going faster than others, just that some thing doesn't smell right in the numbers)

Haggar said...

>>>>>

What time period is the Garmin display in ??
Example being is my last sailing session with GT31 gave me a higher 1 sec average then 2 second. Thats going to always happen I guess.

Otherwise when you post just specify what GPS you have used and if you used trackpoints or doppler.

That's one of the reasons Gramins are less accurate, they only sample every 2secs, so garmin display is the 2sec average, in doppler.
Unfortunately the garmin display readout is doppler, but it doesn't record the doppler data. That's why the display is more accurate.

Hi Paul,

The problem with people using trackpoints is that most people don't know what a spike looks like.

They load up their file, type in 6 numbers and a quick story and hit submit.

If you are experienced in data analysis then this isn't an issue, but most people aren't experienced (and sometimes don't even click on each value to see what it looks like on the map display).

This is why I would prefer that everyone use doppler (or be penalised for using trackpoints). You don't have to be a rocket scientist to figure out how fast you went if you use doppler.


Some people wrote above that the team aspect will average out any high scores, but this doesn't take into account the personal rankings -- which some people put a lot of weight on. What use is a personal ranking page if you can't trust the data that's on it?


I guess the general consensus is that we shouldn't change the numbers that people submit, but we should put in an option to say which device and calculation method the person is using. This would then make it obvious how much faith we can put on that number.
I would also like to put something in the rules to say that you "enter the speed that you think is correct". If you choose to use claimed speed, doppler or trackpoint, then that's up to you.


It might seem to some people that I'm being overly pedantic, but people really do take this thing seriously, and if we can get the rules correct then we hopefully won't have another fiasco like as happened this year where someone decided to leave the challenge. I really don't want that to happen again.

nebbian said...

Hi Paul,

The problem with people using trackpoints is that most people don't know what a spike looks like.

They load up their file, type in 6 numbers and a quick story and hit submit.

Won't they continue to do that anyway?
I thought you could set the GTs to either Doppler or Trackpoint.
Would using Trackpoint induce the risk of spikes with the GTs?

If you are experienced in data analysis then this isn't an issue, but most people aren't experienced (and sometimes don't even click on each value to see what it looks like on the map display).

That's a worry because I discard a lot of Speeds & Alphas because of spikes, maybe more education is required.

This is why I would prefer that everyone use doppler (or be penalised for using trackpoints). You don't have to be a rocket scientist to figure out how fast you went if you use doppler.

Are the tracks that much cleaner using Doppler?

Some people wrote above that the team aspect will average out any high scores, but this doesn't take into account the personal rankings -- which some people put a lot of weight on. What use is a personal ranking page if you can't trust the data that's on it?

It would seem we can only go by trust anyway as we upload our own numbers

I guess the general consensus is that we shouldn't change the numbers that people submit, but we should put in an option to say which device and calculation method the person is using. This would then make it obvious how much faith we can put on that number.
I would also like to put something in the rules to say that you "enter the speed that you think is correct". If you choose to use claimed speed, doppler or trackpoint, then that's up to you.

If we can't all downgrade to Trackpoint & it's unreasonable to demand replacing GPI then that's probably the best idea

It might seem to some people that I'm being overly pedantic, but people really do take this thing seriously, and if we can get the rules correct then we hopefully won't have another fiasco like as happened this year where someone decided to leave the challenge. I really don't want that to happen again.

I'm just trying to get a better understanding of the differences & BTW I am changing to a GT31

nebbian said...

Hi Paul,

The problem with people using trackpoints is that most people don't know what a spike looks like.

They load up their file, type in 6 numbers and a quick story and hit submit.

This is not really a Trackpoints issue. It is really a GARMIN issue. If you use trackpoints from the GT-11/31 you will have the info in the files for the automatic filters in Realspeed and GPS-Results to filter out 99.9% of errors (spikes), just as those programs do with Doppler data.

nebbian said...


This is why I would prefer that everyone use Doppler (or be penalised for using trackpoints). You don't have to be a rocket scientist to figure out how fast you went if you use Doppler.

What we really need is everyone to use GT-11/31. After that it is far less important if we do the rankings on Doppler or Trackpoints

nebbian said...


Some people wrote above that the team aspect will average out any high scores, but this doesn't take into account the personal rankings -- which some people put a lot of weight on. What use is a personal ranking page if you can't trust the data that's on it?

This primarily a Team Challenge. The individual rankings are just there for.... well... I really don't know. If you are into individual rankings, that is what GPS-SS is for and there Doppler rules. (or it will very soon )

nebbian said...



I guess the general consensus is that we shouldn't change the numbers that people submit, but we should put in an option to say which device and calculation method the person is using. This would then make it obvious how much faith we can put on that number.
I would also like to put something in the rules to say that you "enter the speed that you think is correct". If you choose to use claimed speed, doppler or trackpoint, then that's up to you.

Can't see much point in that rule. That is pretty much the defacto situation. It is up to you, but there is still a good reason for teams to post trackpont results as long as there are Garmin users posting without correction.

nebbian said...



It might seem to some people that I'm being overly pedantic, but people really do take this thing seriously, and if we can get the rules correct then we hopefully won't have another fiasco like as happened this year where someone decided to leave the challenge. I really don't want that to happen again.

Not at all Ben. It is very important to have this discussion so people understand and can weigh up the implications and to try to avoid further misunderstandings.

At some time we are going to have to bite the bullet and penalise any remaining Garmin users. Even in 2010 there is sure to be a few still hanging around. Might as well go to 'Claimed Speed' in 2009 and get it over with. If people are just in it for 'a bit of fun' it shouldn't worry them too much. If they are a bit more serious then it is great incentive to upgrade their GPS.